2024-08-29
3726
#nextjs
Peter Ekene Eze
183861
Aug 29, 2024 ⋅ 13 min read

Diving into Server Actions in Next.js 14

Peter Ekene Eze Learn, Apply, Share

Recent posts:

react children prop how to properly type

How to type React children correctly in TypeScript

Learn modern best practices for typing React children in TypeScript, including ReactNode, PropsWithChildren, ComponentProps, and why React.FC is no longer recommended.

Ohans Emmanuel
Dec 19, 2025 ⋅ 7 min read
complete guide to internationalization next js

The complete guide to internationalization in Next.js

Learn how to internationalize Next.js apps with Lingui and next-intl, covering App Router, RSC, routing, locale detection, and dynamic language switching.

Ivan Vlatkovic
Dec 19, 2025 ⋅ 13 min read

Vite vs. Webpack for react apps in 2025: A senior engineer’s perspective

Vite vs Webpack in 2025: a senior engineer’s take on performance, developer experience, build control, and when each tool makes sense for React apps.

Peter Aideloje
Dec 19, 2025 ⋅ 3 min read
vitest 4 adoption guide

Vitest 4 adoption guide: Overview and migrating from Jest

Learn how Vitest 4 makes migrating from Jest painless, with codemods, faster tests, native ESM, browser testing, and a better DX.

Onuorah Bonaventure
Dec 18, 2025 ⋅ 15 min read
View all posts

2 Replies to "Diving into Server Actions in Next.js 14"

  1. So the main issue was that if we posted a new item into a list, we had to do the POST request and then pull data again with a GET to have the UI updated? And this was expensive because there are 2 requests for 1 action, basically, right?

    What we did was:
    – we had a layer for state management (i.e Redux)
    – a layer for API calls (middleware on client side)

    Upon a request, we updated UI optimistically or pessimistically, depending on what the team wanted.

    So, when someone does a POST request, it calls the middleware API, We await the status code, if it’s 200, we update the UI (redux state) and this way, we achieve consistency between server and client’s UI in 1 request, not 2. Which means only 1 server is needed (some S3 bucket for the react site files) and 1 action needed (like the pessimistically update mentioned above). No need for having a second server running the server-side components, no need to trigger a GETall after a POST to see the updated UI. Nothing. Just a well structured and organized application. And it still adheres to separation of concerns since on the backend I run edge functions for the updates etc. It’s just more things on the client. And I am totally fine with SSG because yes, first load is larger but then everything else is a breeze, not requiring calls to a server. But yeah, I guess anything goes nowadays.

Leave a Reply

Hey there, want to help make our blog better?

Join LogRocket’s Content Advisory Board. You’ll help inform the type of content we create and get access to exclusive meetups, social accreditation, and swag.

Sign up now